Bartosz Marcinkowski

Bio

Early in my career I enjoyed developing software for data processing. Over time, I developed interest in machine learning and data science. In parallel, progressing wrists issues pushed me away from an implementation-heavy path towards research, management and business-oriented work.

Today I enjoy a mix of applied ML research, using data science to achieve business goals, managing projects and mentoring.

Work experience

2023-now Senior Data Scientist / Team Leader, MIM Solutions, remotely

I lead projects in

2021-2023 Machine Learning Researcher, RTB House, remotely

I lead ML development of a new product and coordinated related cross-team efforts. I also managed and mentored junior ML researchers.

2021 Machine Learning Engineer / Researcher, RTB House, remotely

I worked on a mix of engineering tasks (developing an ML pipeline) and research tasks (developing models predicting internet users’ behavior, future transaction value and providing recommendations).

2016-2020 Senior Software Engineer, dimensions.ai, remotely

I worked on text classification, term extraction, named entity recognition and active learning.

2014-2016 Software Engineer, RTBHouse, Warsaw, Poland

I was a part of the engineering core of a Real-Time Bidding platform backed by the Hadoop ecosystem.

2013-2014 Software Engineer, SzybkaPapuga.pl, Warsaw, Poland

I worked on optimizing database interaction of a legislation searching system.

Projects

2024 Towards Semantic Embeddings of Cardiological Signals with Diffusion Autoencoders https://doi.org/10.36227/techrxiv.173152781.12841872/v1

I developed a diffusion autoencoder to support the development of wearable medical devices for remote monitoring and treatment of cardiovascular diseases. The work was a collaboration between MIM Solutions and CSW Therapeutics.

2024 flow matching notebook, https://bm371613.github.io/conditional-flow-matching/

I created this notebook while working at MIM Solutions on a project where we developed diffusion and flow matching based models for self-supervised representation learning of ECG signals. I wanted to convey the basic intuition with visualizations.

2023 MIRAD, https://github.com/Sagenso/MIRAD/

MIRAD is a machine learning project done for a cyber-security company Sagenso that wanted to detect ransomware attacks. The company collected system logs during simulated attack scenarios and needed an intepretable model that would detect attacks, but also provide insights that could lead to improvements of the simulations.

2020 gest, https://github.com/bm371613/gest

Gest allows using hand gestures as an input device. I used it on daily basis alongside Talon (speech recognition, eye tracking) as a wrist-friendly replacement for a mouse and a keyboard.

Education

2014-2017 Master’s degree in Computer Science, Warsaw University

I picked my courses to form two parallel paths: algorithms/mathematics/machine learning and system programming/data processing.

My thesis was on implementing artificial intelligence for a board game.

2011-2014 Bachelor’s degree in Computer Science, Warsaw University

I was initially focused on algorithms and mathematics, and developed a parallel interest in system programming over time.

Most of this programme I received a scholarship based on my grades.

Honours

2011 Finalist, Polish Olympiad in Informatics

2011 Finalist, Polish Olympiad in Mathematics

2010 Finalist, Polish Olympiad in Mathematics

Certifications

2010 Certificate in Advanced English, Cambridge Assessment English

Grade A certifies C2 fluency.